Pulse-Code Modulation Technique for Digital Telephone

.•

1911 NEAR EAST TECHNICAL UNIVERSI1Y ELECTRIC AND ELEC1RONIC ENGINEERING DEPARTMENT

PULSE-CODE MODULATION TECHNIQUE FOR DIGITAL TELEPHONE

SUPERVISOR

Prof.Dr. Fahretttin SADIKOGLU MAMEDOV ACKNOWLEDGEMENTS

The authors would like to thank Prof. Dr. Fahrettin M. Sadikoglu for his supervision and constructive criticism in the development of this thesis. The authors also wishes to express his deep gratitude to his par• ents for their support throughout his life. • \ TABLE OF CONTENTS ,) /

INTRODUCTION . 1 CHAPTER I : THE FUNCTION BLOCK OF THE PCM SYSTEM . 4 1.1. Sampling . 6 1.2. Quantizing; . 8

1.3. Coding & Companding . 9 1.4. Electrical Representation of PCM Signals 16 1.5. Demodulation 18 CHAPTER II: PCM TRANSMISSION SYSTEM 21 2.1. First Order PCM Systems 21 2.2. First Order PCM System Recommended by CCITI Frame Structures 23

2.3. Second Order PCM systems 28 2.4. tdm Hierarchy 29 CHAPTER III: THE TELEPHONE NEIWORK 36 3.1. The Public Switched Telephone Network (PSTN) 36 3.1.1. Switching Hierarchy of North America 38 3.2. Technical Considerations In The Planning of Digital Network For Multiexchange Areas 40

3.2.1. Numbering 41 3.2.2. Transmission System 42 3.2.3. Transmission Planning 43 3.2.4. Signalling 44 3.3. Network Structure In The Rural Area 46 3.4. A Completely Digital Network 4 7

CONCLUSION 50 REFERENCES .•

INTRODUCTION

Pulse - code modulation (PCM)was developed in 1937 at the Paris Laboratories of American Telephone and Telegraph Company (AT&T) by Aled Reeves. Reeves conducted several succesful transmission experi• ments across the English channel using the modulation techniqe, in• cluding pulse-width modulation (PWM). pulse-amplitude modulation (PAM) and pulse-position modulation (PPM).At the time circuits involved was enormously complex & expensive. Although the significance of Reeve's experiments was acknowledged by Bell laboratories, it was no until the semiconductor industry evolved in the 1960s that PCM because more prevalent. Currently in the world PCM is the prefered method of trans• mission within the Public Swiched Telephone Network (PSTN). PCM is a methode of digital transmission of an analog signal. The motives for using PCM in the telephone network can be summarized in the following7 points: 0 Transmission quality almost independent of distance. A char• acteristic of a digital signal is its immunity to interference. Digital signals can be regenerated at intermediate points along a transmission line with• out loss of quality. This is not the case with analogue signals where not only the signal but also the noise is amplified at the intermediate am• plification points. In the digital case the repeaters have to make the simple decision as to whether an incoming pulse is a "one" or a "zero". After the decision has been taken a fresh pulse is transmitted. It is certain that a number of incoming pulses will be so distorted that they cannot be rec• ognized correctly, but this failure rate can be made as low as required. It should however be observed, that the PCM-systems used in practice and

1 specified by CCITT are not, from a transmission quality point of view bet• ter than the FDM-system specified by CCITT. 0 Time-division-multiplex. The TDM principle allows an increase in capacity on cable pairs originally used for single telephone channels. It may be feasible to introduce PCM transmission on these pairs instead of laying new cables. 0 Economy for certain links. In certain applications, especially in the junction network, PCM transmission has shown itself to be com• petitive with any other method of transmission. The length of the trans• mission links should be in the inttermediate region where normal voice frequency (VF) links tend to be too long and FDM links too short. This optimal distance interval is very dependent on local factors such as subscriber density, topology of the country etc .. and therefore var• ies widely. However, the figures in the graph can be taken as being typical for the first order PCM multiplex system. O Economy in combination with digital switching. A high proportion of the cost for PCM-systems lies in the terminal cost. The introduction of digital switching will lsubstanttally reduce this cost since switching is per• formed directly on the digital bit stream and no costly analogue/ digital conversion is needed. A combination of digital switching and transmission will therefore tend to lower the overall cost. 0 IC-technology. Developments in integrated circuit technology seem to point to favourable cost levels and a high degree of reliabilty.

0 Integration of services. As a digital medium a PCM-link can transmit not only speech but also data, telex, encoded visual information etc. A PCM-channel has a capacity of 64,000 bit/s which makes it a very powerful data channel.

2 •

New transmission media. Future wide-band transmission media, such as waveguides and optical fibres are more suitable for digital than analogue transmission.

The project consists of 3 chapters:

Chapter 1 presents five major principles: sampling, quantizing, electrical representation of PCM signals, Coding and Companding, de• modulation for PCM systems. This theoretical material is used it be repre• sentation first order PCM system. The functional and timing diagram is system are presented.

Chapter 2 starts with description 30 and 24 channels standard PCM systems rdecommended by CCIT.The more important accept of PCM modulation technique-digital multiplexing hierarchy is analysed. The eval• uation of Transmission rate for different level hierarchy are clearly cal• culated and explained. Chapter 3 includes an application PCM system in the public switched telephone network. Different types of network is considered. New method of switching is given to interconnect ceutial office through the use of interoffice trunks and tudem trunus. Outside the local area told trunus and intertool trunus are considered. The text is arrauged so that it can be used by students to study same part of the subject "Telecommunication". It can be useful for en• gineers too.

3 CHAPTER I THE FUNCTION BLOCK OF THE PCM SYSTEM

The statement above gives some idea about the basic processes in pulse code modulation. Here we shall give these processes their right names. The process of choosing measuring points on the analoguee speech curve is called sampling. The measurement values are called samples. When samlpling, we take the first step towards a digital representation of the speech signal as the chosen sampling instants give us the time co• ordinates of the measuring points. The amplitudes of the samples can assume each value in the am• plitude range of the speech signal. When measuring the sample am• plitudes we have to round off for practical reasons. In the rounding-off process, or the quantizing process, all sample amplitudes between two marks on the scale will be given the same quantized value. The number of quantized samples is discrete as we have only a discrete number of marks on our scale. Each quantized sample is then represented by the number of the scale mark, i.e. we know now the coordinates on the amplitude axis of the samples. The processes of sampling and quantizing yield a digital repre• sentation of the original speech signal, but not in a form best suited to transmission over a line or radio path. Translation to a different form of signal is required. This process is known as encoding. Most often the sam• ple values are encoded to binary form, so that each sample value is repre-

4 .. sented by a group of binary elements. Typically, a quantized sample can assume one of 256 values. In binary form the sample will be represented by a group of 8 elements. This group is in the followingcalled a PCMword.

For transmission purposes the binary values O and I can be taken as cor• responding to the absence and presence of an electrical pulse. On the transmission line the pulses in the PCM words will become gradually more distorted. However, as long as it is possible to distinguish between the absence and the presence of a pulse, no information loss has occurred. If the pulse train is regenerated, i.e. badly distorted pulses are replaced by fresh pulses at suitable intervals, the information can be transmitted long distances with practically no distortion at all. This is one of the advantages of digital transmission over analogue transmission; the information is contained in the existence or not of a pulse rather than in the form of the pulse. In our picture of the graph and the table this is analogous to the fact that the information in the table is not affected if the digits are badly written as long as they are legible. But if the graph is badly drawn, loss of information is inevitable. On the receiving side the PCM words are decoded. t.e. they are translated back to quantized samples. The analogue speech signal is then reconstructed by interpolation between the quantized samples. There is a small difference between the analogue speech signal on the receiving side and the corresponding signal on the transmitting side due to the rounding off of the speech samples. This difference is known as quantizing dis• tortion. The function blocks in the pulse code modulation process are shown in figure 1. 1.

5 Analogue signal

Receiver

Transmlsslon line PCM signal

Figure 1 - 1. Pulse code modulation Function blocks

1.1. Sampling

In the practical electrical meaning, to sample is to take instaneous values of the analogue signal at equal time intervals. See figure 1-2.

Orlqlnal fA '~".'' 1 ~--=:;/ •o,m, Sampler , Amplitude Samples

Sampled ~/J s~~~---- I I I I "'lime - Sampling Interval t, ' f • J_ Sampling rate ' t,

Figure 1 - 2. The sampling process

The sampled signal is a train of pulses, whose envelope is the orig• inal signal. Now, what should be thesampling rate, i.e. the number of samplles per second? The answer to this question is given by the Sampling Theo-

6 rem, which also illustrates the fundamental fact that the information con• tained in the signal is not affected by sampling: The sampled signal contains within it all information about the original signal if: 0 the original signal is band limited, i.e. it has no frequency

components in its spectrum beyond some frequency B

0 the sampling rate is equal to or greater than twice B, i.e. fs ~ 2B. The sampling theorem is illustrated in figure I-3. Obviously, the spectrum of the sampled signal contains the spectrum of the original signal, i.e. no information loss has occurred. _c_JCC •. frequency a) band limited signal

___ :~b'"nh ch l I B Is 21s frequency b) sampled signal, Is> 2 · B

Figure 1 - 3 . Spectrum of a) band-limited signal b) sampled signal

In telephony, the part of the speech spectrum between 300 and 3400 Hz is used. The human speech spectrum extends from a lowest fre• quency of some 100 Hz up to very high audio frequencies. The telephone set reduces this frequency range, but not enough at high frequencies so in order to come below this band limit at 3400 Hz, the speech signal must be low-pass-filtered before sampling.

7 .• plitude. This means that a loud talker and quiet talker let a listener hear the same quantizing distortion. Relative to the speech levels, the quiet talker generates much more distortion than the loud talker. Furthermore, a statistical analysis shows that for an individual talker, small amplitudes are much more probable than large ones.

amplitude • ,

sampled signal (PAM signal)

quantizer

quanllzed signal I quantizing Interval )------quantizing level

Figure I - 4 . The quantizing process

1.3. Coding & Companding Practical PC systems use seven-and eight-levelbinary codes, or

27 = 128 quantum steps

28 = 256 quantum steps Two methods are used to reduce the quantum steps to 128 or 256 without sacrificing fidelity. These are nonuniform quantizing steps and companding before quantizing, followed by uniform quantizing. unlike data transmission, in speech transmission there is a much greater like• lihood of encountering signals of small amplitudes than those of larger amplitudes. A secondary but equally important aspect is that coded signals are designed to convey maximum information, considering that all quantium steps (meanings or characters) will have an equally probable occurrence

9 A sampling rate of 8000 Hz is used for PCM systems in telephony. This rate is somewhat higher than twice the highest frequency in the band, 3400 Hz, due to difficulties in making low-pass filters steep enough. The sampled signal is often said to be pulse amplitude modulated as it consists of a train of pulses, whose amplitudes have been modulated by the original signal. Pulse Amplitude Modulation (PAM) is an analogue pulse modulation method as the amplitudes of the pulses may vary con• tinously in accordance with the original signal variations. The relative simplicity of PAM systems makes them attractive for some telephony applications. However, PAMis unsuitable for transmission over long distances owing to the difficulty of pulse regeneration with suf• ficient accuracy, which is important as the PAM pulses contain the in• formation, in the pulse form.

1. 2. Quantizing

The continuous pulse amplitude range is broken down to a finite number of amplitude values in the quantizing process. The aplitude values in the quantizing process. The amplitude range is divided into intervals, and all samples whose amplitudes fall into one specific quantizing interval are given the same output amplitude. See figure I-4. The rounding off of the samples causes an irretrievable error, quanting distortion, in the sig• nal. This voluntary sacrifice, which can be brought down to suitable low limits by making the number of permitted amplitude levels large enough, is accepted because it makes error-free transmission possible by only having a discrete number of amplitudes. In figure I-4 the quantizing distortion is independent of sample am-

8 ..

(i.e., the signal-level amplitude is assumed to follow a uniform probabuty distribution between O and ± the maximum voltage of the channel). To cir• cumvent the problem of nonequiprobability of signal level for voice signals, specifically, that lower - level signals are more probable than higher-level signals, larger quantum steps are used for the larger-amplitude portion of the signal, and finer steps are used for the signals with low amplitudes. The two methods of reducing the total number of quantum steps can now be more precisely labeled: * Nonuiform quantizing performed in the coding process. * Companding (compression) before the signals enter the coder, which now performs uniform quanttizing on the resulting .stgnal before

coding. At the receive end, expansion is carried out after decoding. Most practical PCM systems use complanding to give finer gran•

ularity (more stepss) to the smaller amplitude signals. This is in• stantaneous companding, as compared to the syllabic companding used in analog carrier telephony. Compression imparts more gain to lower am• plitude signals. The compression and later expansion functions are log• arithmic and follow one of two laws, the A law or the "mu" (µ)law. The curve for the A law may be plotted from the formula Alxl ) ( 1 + ln(A) x ~ lxl ~ l 1 + Inlxl ) 1 A ( 1 + ln(A) A s lxJ ~ 1

where A= 87.6. The curve for theµ law may be plotted from the formula:

y = ln ( 1 + µJxl ln (1 + µ)

where x is signal imput amplitude and µ = 100 for the original North American Tl system (now out dated) and 255 for later North American

(DSI) systems and the CCITI 24-channel system.

10 A common expression used in dealing with the "quality" of aPCM signal is the signal-to-distortion ratio (expressed in decibels). Parameters A andµ determine the range over which the signal-to-distortion ratio is com• paratively constant. This is the dynamic range. Using aµ of 100 can pro• vide a dynamic range of 40 dB of relative linearity in the signal-to• distortion ratio. In actual PCM systems the companding circuitry does not provide an exact replica of the logarithmic curves shown. The circuitry produces approximate equivalents using a segmented curve, and each segment is linear. The more segments the curve has, the more it approaches the true logarithmic curve desired. Such a segmented curve is shown in Figure 1-5. If µ law were implemented using a seven (hetghtj-segment linear ap• proximate eqivalent, it would appear as shown in Figure 1-5. Thus on coding, the first three coded digits would indicate the segment number

(e.g. 23 = 8). Of the seven-digit code, the remaining four digits would di• vide each segment into 16 equal parts to identify further the exact quan• tum step (e.g., 24 = 16). For small signals, the companding improvement

1.0

~ ci"' E 0 u

'-::---'--__1_ _ _.l_-f _j__J -1.0 Input 1.0

Figure I - 5. Seven-segmentlinear approximate of the logarithmic

curvefor µlaw(µ= 100)

11 is approximately

A law: 24 dB µlaw: 30 dB using a seven-level code. These values derive from the equation of com• panding improvement or

Gc!B = Uniform (linear) scale Companded scaJe

Coding in PCM systems utilizes straighforward binary codes. Examples of such coding are shown in Figure l-5a, which is expanded in Figure 9.7, and in Figure 9.8, which is expanded in Figure 1.6.b showing a number of example code levels.. The coding process is closely related to quantizing. In practical sys• tems, whether the A law or the µ law is used, quantizing employes seg• mented equivalents of the companding curve (Figures I-6 andl-8), as dis• cussed earlier. Such segmenting is a handy aid to coding. Consider the European 30 + 2 PCM system, which uses a 13-segment approximation of the A-law curve (Figure 1-6). The first code element indicates whether the quantum step is in the negative or positive half of the curve. For example, if the first code element were a 1, it would indicate a positive value (e.g., the quantum step is located above the origin). The followingthree-code ele• ments (bits) identify the segment, as there are seven segments above and seven segments below the origin (horizontal axis). The first four elements of the fourth + segment are 1101. The first 1 indicates it is above the horizontal axis (e.g., it is positive). The next

12 three elements indicate the fourth step or

0-1000 and 1001 1-1010 2-1011 ~3-1100 4 -1101 5-1110 etc.

Corle Segment 1 1 1 1 X XX X I ____0 ,,2 --·- I 1 1 1 0 X X X X -~ I OG I 1101xxxx I 80 2 I 3 ,ooxxxx 4 I 64 I 1011xxxx 2 48 _L1v1 I J 32 101oxxxx I 32 4 J. (VI I 64 1001xxxx I (VI 1 £ (VI ltVI I 0 I 4tVI 4 4 1 oooxxxx l IL__j l 1 0 ooooxxxx 0001xxxx

001oxxxx

0011xxxx

01ooxxxx

0101xxxx

011oxxxx

0111xxxx

Figure I - 6. Quantization and coding used in the CEPT 30 + 2 PCM system.

13 .•

Figure 1. 7 shows a "blowup" of the uniform quantizing and sub• sequent straightforward binary coding of step 4. This is the final segment coding, the last four bits of a PCM code word for this system. Note the 16 steps in the segment, which are uniform in size.

110101()()<----

Figure I - 7. The CEPT30 + 2 PCM system, coding of segment (4 positive).

Encoding Code

1 3 5 6 7 8 Segments Encoder input 16 Steps each

Figure I - 8. Positive portion of segmented approximation of u law quantizing curve used in North American (AIT) DS1 PCM channelizing equip• ment. Courtesy of IIT Telecommunications, Raleigh, N.C.

14 The North American DSI PCM system uses a 15-segment ap• proximation of the logarithmic µ law. Again, there ase actually 16 seg• ments. The segments cutting the origin are colinear and counted as one. The quantization in the DSI system is shown in Figure 1-8 for the positive portion of the curve. Segment 5, representing quantizing steps 64 through 80, is shown blown up in Figure 1-8, Figure 1-9 shows the DSI coding. As can be seen in the figure, again the first code element, whether a 1 or a 0, indicates whether the quantum step is above or below the horizontal axis.

The next three elements identify the segment, and the last four elements (bits) identify the actual quantum level inside the segment. Of course, we see that the DSI is a basic 24-channel system using eight-level coding with

µ-law quantization characteristic whereµ= 255.

l>igi1 N 11111her

:> 7 8 Code I .evvl I 2 '.I 4 6

() () () () () () 0 255 (l't'ak posili\'c le\'l'I) I () () () 0 23!1 I () () I () () () () () 223 I () I () () () () 207 I () I I () () () () () 11!11 I I () ) () () () () () ',17:> I I I () () () () 0 ):>!I I I I I () () 0 () 14'.1 I I I I I I 127 (( .euu-r levels) I I I I I 12fi ( Nominal tt'ro) () () () 0 () 111 () I I I () () () () 9:> () I I () () () ;( 79 () I () I () () () () 63 () I () () () 0 () () (l () 47 () () I I () () () () () 31 () () I () () () () 15 () () () I () () 2 () () () () I I () () () () () () I () () () () 0 (l't'ak lll'gati\'t' h-vr-l) () 0 () () I* *One digit is ad,ll-d to eusurc that rl«: 1i111ing conte111 of the uansruiued paltt'rn is mainraiucr], · Figure I - 9. Eight-level coding of North American (ATT) DS1 PCM

system. Note that there are actually only 255 quantizing steps because steps O and 1 use the same bit sequence, thus avoiding a code sequence

with no transitions ii.e., O's only).

15 As we know, pulses with two levels, i.e. binary pulses, are at• tractive for transmission as they are easy to regenerate on the trans• mission line. It is not dfficult to build regenerator circuits able to de• termine whether a pulse is present or not. Present-day practical system use binary encoding of the quantized speech samples. See figure 10. As telephony uses 256 quantizing levels, each sample will be encoded to a code group. or PCM word. consisting of 8 binary pulses (8 bits).

amplllude

111 ++3

110 + +2 quantized 100 .o, I I I I ... , .. 000"' r-0 I 1 , , I 001 -1 I . I 010 -21

Encoder I I 011 + -3 I I I

pulse code modulated 100110111101001010000101.....__.....,, ....._,__.. ...____,_, ...... ,_._ '--....--' signal ..____. ____, ____.., +O +2 +3 +1 -1 -2 -0 +1

Figure 1- 10. Encoding of quantized samples with 8 quantizing lev•

els (3 binary digits/code word).

As the sampling rate used is 8000 samples/second, one pulse code modulated speech signal will generate a 64 kbit/s digital signal.

1.4. Electrical Representation of PCM signals

Digital signals within the terminal are usually transmitted in the form of a unipolar pulse train in the nonreturn-to-zero (NRZ) mode, see figure 11. This signal form is not appropriate for transmission over long distances.

16 .•

I I IO I 0 I I ]( I r

Figure 1- 11. Binary information represented in:

I a unipolar nonreturn-to-zero (NRZ) pulse train.

II · a bipolar return-to-zero (RZ) pulse train.

A better form is a bipolar return-to-zero (RZ) signal. The ad• vanteges of this signal are : * it has no power in the lower parts of its spectrum, i.e. it has no direct current component; this is due to the alternating polarities of the pulses * the intersymbol interference is reduced by the return-to-zero feature. Of course, even this signal will be attenuated and distorted during transmission, and noise will be added to It. At some point on the transmission line the signal must be restored. This is done by inserting a device on the line that first examines the dis• torted pulse train to see whether the likely binary value is 1 or 0, and then generates and transmits to the line new pulses according to the result of the examination. Such a device is called a regenerative repeater. See figure 12. At the same time as the pulses are reshaped, the noise added during transmission is eliminated at the least if the noise signal amplitude is not large enough to bring the received code signal to the wrong side of a re• generator decision level. Normally, the regenerated code signal is identical to the transmitted original code signal. Even after a large number of re-

17 generative repeaters, the code signal is practically identical to the original signal. This is the reason for the high transmission quality that is obtain• able with PCM transmission system.

PCM 1,ansm111., I ~ .~n·~· ··~·~~ . l ".1'.!'~·~11~·~"._0~·- ---ir ~ _

, <""", - ,...... ,., d 11 J ,:s;..=: / \ ::Z , 0 - ocson \.T \...? - levels 1 Regenor~llve timing l I t I t I f I repeater ------"" I

: rogeneraledcode ~

Figure 12. Pulse forms on a transmission line.

1.5. Demodulation

The processes in the receiver that convert the incoming PCM signal to an analogue speech signal again are regeneration, decoding and re• construction. The regeneration process has the same aim and is performed in the same way as on the transmission line, i.e. the distorted pulses are replaced by new square pulses, see figure 12. Beforeentering the decoder the bipolar signal is reconverted to unipolar. In the decoding process the code words generate amplitude pulses, whose hetghtts are the same as the heights of the quantized samples, which generated the code words. So after passing through the de• coder the train of quantized samples is retrieved. See figure 13.

18 •.

• 0 +2 .• 3 + 1 . t -2 •O + 1 ,_..,...____,___,...__.__,--..,,..._.....---..,.-__,,--._..--,,,...... _~ pulse code modulated slgnal 1001101111010010 1000010\

amplitude

111 +•J decoder

110 +2

101 + 1 100 +O 1 quantized 000 1 samples 001 -1 time 010 -2

011 + -3

Figure 13. Decoding of encoded amplitude levels

The analogue signal is reconstructed in a low pass filter. figure 14a. This can be seen from figure 14b. The spectrum of a sampled signal contains the spectrum of the original signal as has been shown in figure I- 3. A low pass filter with a cut-off frequency at B Hz takes away all fre• quency components in the spectrum above B Hz and the spectrum of the desired analogue signal is left.

1 amplltuda quantized samples • I I I I I , I I ., time

low pass JIiter

reconstructed ~~~o~~ _s!_9~1 !~'=7•llmo

Figure I - 14.a Reconstruction of the analogue

19 , .. TQ"'"dc\ B Is oh21 l "'"' I :'~T''" ·rrnm • '"' aarnplad ''""" • """'"" -,---1 ~/////fl@/////ft//Pu//$/////4, Power /lrequency low pass tiller characlerlsllc frequency

analogue L-, signal LJ ------~1sr------trequency spectrum of reconstructed analogue signal ••

Figure I - 14b. Reconstruction of the analogue signal shown by spectrum diagram.

20 CHAPTER II PCM TRANSMISSION SYSTEM

We have now dealt with the fundamental principles of PCM. In this part of the chapter we shall describe how these principles are used to build up practical PCM transmission systems. However, we shall start by explaining the time-division multiplex principle as this makes PCM trans• mission systems for telephony economic?-11Yattractive.

2.1. First Order PCM Systems

Several signals in pulse from can use a common transmission path if the signals have different phases. Figure 2-1 shows how three PAM-signalsare time-division multiplexed on the same transmission line. The pulses of the three signals are interleaved by opening the sampling gates one by one cyclically. During one cycle the transmission line receives one PAMpulse from each of the participating signals. Such a set of pulses is called one frame. The time interval that each of these pulses occupies is called a time slot. In this example each frame has three time slots. On the receiving side the pulses are distributed again. This is done by opening the sampling gates cyclically in the same manner as on the transmitting side. Due regard must of course be paid to the transmission delay. This delay has been omitted in figure 2-1 for clarity. In the case of PCM signals the time-division multiplexing is most ofter carried out before the samples are pulse coded, i.e., the samples from the participating analogue signals are combined on a common PAMtrans• mission line. See figure 2-2.

21 .•

ifb- __ I i I ({o- __ 4 VF~ 1...... :: i,. -r ({o- __' : I · · I VF=v~~ ~~I. I I~ P!'~TD"r ~ h p I I I I f==.l : 1trame1 1 J -@- low pass filter 1 1 time I I I 1 ~II -8- sampler P~~T_DMpop.n p.n p :_11 I I I . I : I I I I I I I I I ecoder] I I I 1 I _J I I pl P~~ TDMp I--, P ~ f= I I I I . Gb---\ st: ~ : -~ .•... I I I Cro--- =:::::::::J.. ~ I I .. Gn------VF =7 ~ . "-=..

Figure 2-1. A PAM transmission system using time-division multiplex(TDM)

@ low pass lllter H sampler transmission line'

I l -, I I I

Figure 2-2. A PCM-TDM transmission system. Attenuation and de• lay on the PAM and PCM transmission lines are not shown.

22 ..

In this way the coding equipment can be used in time-division mul• tiplex. We see from the figure that the PCMpulses are not nterleaved pulse by pulse, but PCMword by PCMword. This is often called time slot inter• leaving. PCM systems used in telephony are most often TDM systems, so \ when we read or hear the term "PCMsystem" it is almost always referring to a PCM-TDMsystem. Howeverit must not be forgotten that PCM in itself can be used, and is used in some cases, on a one-channel basis.

2.2. First Order PCM System Recommended by CCITTFrame Structures

As mentioned in chapter I, CCITTrecommends two ftrst order, or primary, PCM systems for use in telephony: the 30-channel system, pro• posed by CEPT, and the 24-channel system proposed by AT & T. The first order systems will form the basis for hierarchies of digital transmission systems. We have to distinguish between the PCM multiplex equipment, or the PCM terminal, and the/PCM transmission line. The multiplex equip• ment converts a number of analogue signals (30 or 24) to a digital signal on the transmitting side and carries out the inverse functions on the re• ceiving side. The transmission line conveys the digital signals between two multiplex equipment units. See figure 2-3. In the following,the 30-channel multiplex equipment will be treat• ed in some detail, as this multiplex forms the basis for the subsequent presentation of digital telephony in this book. A summary of the most im• portant data on the 24-channel multiplex is also given. The presentation of first order PCM systems will end with a brief description of the transmission lines.

23 ~tr~3~ t., ~~-\.•... ~ 11, Il ,, "• \\\

R \ ,--Re-c-ei-ve-r~ ', -H Transmitter -- ·-4- • - i) ..,. _

Figure 2-3. Two multiplex equipment units.

Thirty analogue speech channels together with. associated sig• nalling are converted to one digital signal by the 30-channel system. The structure of this digital signal is shown in figure 2-4.

_J Time i Primary multiplex uooo Ir arnes/s 204a kblt/s

·O 1 i Frame ! 256 bits' Time slot 125µs \ i 8. bits Sync. ch: synchronization channel 3.9 µs Si;in. ch: signalling channel 10011101

Figure 2-4. Frame structure of the 30-channel primary multiplex.

24 The digital signal is divided into frames, with a repetition rate of 8000 frames/sec. This is of course because the sampling frequency is 8000 Hz and the fact that the frame contains one binary coded sample from each of the analogue signals. Each frame consists of 32 eight bit time slots. Of these, 30 time slots are used for PCM channels and the re• maining two for synchronization and signalling. The PCM channels carry analogue signals within the frequency band 300-3400 Hz, coded according to the A-law, see figure 1-5. The synchronization time slot, time slot O in each frame, contains 8 bits whose purpose is to form a recognition signal to the receiver in order to keep this synchronized to the transmitter so that each PCM channel can be correctly identified. This function is the same as the function in• dicated by the sampling control block in figures 2-1 and 2-2. The signalling time slot, No.16, can be used in many ways. The great signalling capacity, 64 kbit/ s, offers flexibility in chasing suitable schemes for different purposes. This is important when considering the digital network of the future. CCITI has recommended the use of the signalling time slot for ei• ther common channel or channel associated signalling. The arrangements for common channel signalling are not yet specified, so here we can only go into details concerning the channel associated signalling scheme. The signalling scheme is the one used today when introducing PCM primary systems into the existing network. The scheme uses the time slots 16 in sequences of 16 frames, referred to as multiframes, as shown in to figure 2-5.

25 £..::,, I Frame i 0 I I I I -- --- ~---- i__L.__J io 16 31 1 I I I -----~ -- L.__J 0 16 31 2 I I I _____ LLi..!ZJ__ - l__j -I 0 ~-;- 16 31 a, I ·E I £ 14 l..!!..1.2.J l__j E j O 1 16 31 ~ 15 I; I '------~------l__J I O 1 16 31 I multlframlng word 1M:

Figure 2-5. Structure of the channel associated signalling scheme for the 30-channel PCM system.

In the first frame of tthe sequence, frame 0, time slot 16 carries a multiframing word, i.e. a recognition signal that tells the receiver that a new multiframe has started. The eight bits of time slot 16 in the next frame, frame 1, are divided so that the first four bits carry signalling in• formation associated with PCM channel 1, and the last four bits carry sig• nalling information associated with PCM channel 16. In frame 2, time slot 16 carries signalling information associated with PCM channel 16. In frame 2, time slot 16 carries signalling information for the channels 2 and 17 and so on up to frame 15, the last frame in the multiframe, which car• ries signalling information for the channels 15 and 30. Then the next frame is frame O in the next multiframe. Tihus four signalling bits are associated with each PCM channel. Each bit can be used to reproduce the state of a signalling relay in a junc• tor connected to the PCM terminal, i.e. the scheme provides four sig• nalling channels per PCM channel; Normally, for conveying conventional signals, only one or two of the channels are used.

26 .•

The 24-channel PCM multiplex has a somewhat different structure as can be seen from figure 2-6.

PCM channels

23 24 s 2 3 --- 22 s: synchronization bit

Figure 2-6 Frame. structure of the 24-channel system.

No special time slot is assigned to signalling.A channel associated sig• nalling scheme is achievedby taking the least significantbit in every PCMchan- • nel for signallingpurposes every 6th frame. For synchronization one extra bit is inserted in the frame. This bit can also be used for common channel signalling. A summary of important technical data for the primary multiplexes is shown in figure 2-7. It is obvious that the primary multiplexes are not compatible: they have, for example, a different number of time slots and different signalling possibilities. Not even the time slots are compatible as the systems use different encoding laws. However, work is going on at CCITT aimed at finding a digital method for conversion between PCM words using different encoding laws in order to avoid conversions to analogue when connecting different PCM systems to each other.

.the 30-chaJ}llel the 24-channet • , system system audio frequency band 300 -3400 Hz 300 -3400 Hz sampling rate 8000 Hz 8000 Hz bits/sample 8 8* time slots/trame 32 24 PCM channels/frame 30 24 output bit rate 2048 kbit/s 1544 kblt/s encoding law A-law, A=87.6 µ -law, µ =255 signalling capacity

channel associated signalling 11-4 sign. ch./PCM ch. 1-2 sign. ch./PCM ch. common channel signalling · 64 kbit/s I 4 kbit/s •Jonly 7 bns each 6th frame l1 channel associated signalling Is used.

Figure 2-7 Technical datafor the primary multiplexes.

27 The PCM transmission lines used for interconnecting primary mul• tiplexes are most often already existing pair cables used for analogue voice frequency transmission. For a PCM line we need two pairs, one for each di• rection see figure 2-3. The line must be equipped with regenerative re• peaters every 1.5-2.5 km, depending on the type of cable. This distance is about the same as the distance between loading coils on loaded cable cir• cuits, and conversion of loaded pairs into PCM transmission lines can be carried out merely by replacing the loading coils with PCM regenerative re• peaters on specially selected pairs. The transmission lines are designed to convey digital signals of specified rates 2048 or 1544 kbit/s. The dtgttalstgnals must fulfil re• quirements concerning not only pulse polarities, see figure 1-7, but also pulse distribution. The timing circuits in the repeaters need, for their op• eration, digital signals with alternating pulses. Long sequences of zero-bits must be avoided. This is done by a special line code that converts some of the zero-bits to pulses in such a way that these extra bits can be removed before arriving at the decoder.

2.3. Second Order PCMSystem

The primary PCM systems are intended for short-distance ap• plications. In the medium and long distance network, where high chan• nel capacity is demanded, it is more economical and practical to group together larger number of PCM channels to one common transmission line, thus forming higher order systems, than to use several primary PCM systems. In general, multiplexes can be of two types:

28 PCM multiplexes and digital multiplexes

v' PCM multiplexes drive one sngle digital signal from a number of analogue signals by a combination of pulse code modulation and time divi• sion multiplexing and also carry out the inverse function. For example, the first order multiplexes decrtbed earlier are of this type. v' Digital multiplexes derive one single digital signal by combining a number of digital signals by time division multiplexing and also carry out the inverse function. Digital transmission lines convey dgttal signals between multiplex equipment units. These lines are designed to carry digital. signals of spec• ified rates, but they are not dependent on what type of original signal is conveyed in digital form. Tihat is, the same transmission line can be used for both PCM multiplexes and digital multiplexes if these have the same rates on the multiplexed digital signals. Even other types of digital signals, for example digitally coded visual telephone signals or data signals, can use the transmission lines if their digit rates are correct.

2.4. TDM merarchy

Two second order systems have been recommended by CCITI. These have digital multiplexes and are based on each of the two primary multiplexes. They both combine four primary PCM signals to one digital signal. See figure 2-8. The signals are multiplexed by bit interleaving, that is the par• ticipating signals are combined bit by bit, see figure 2-9. This is more practical than time slot interleaving, in which the participatting signals are combined time slot by time slot, as in this latter case it is necessary to col-

29 •

lect the bits of the time slots in buffers before interleaving can be carried out.

~--, Primary __ fl N : PCM I PCM l N ~-- Multiplex Multiplex --~

a) N=30 2048 kbit/s 8448 kbit/s 2048 kbit/s b) N=24 1544 kbit/s 6312 kbit/s 1544 .kbtt/s

Figure 2-8. Second order digital multiplexing of primary PCM sig•

nals. (a) The CEPT scheme. (b) The AT & T scheme.

4------Q

3-----l'rl--

1234112341123411234

Figure 2-9. Bit interleaving

30 ..

The digital multiplexes must accept that the primary signals, for practical reasons, have bit rates slightly differing from the ideal bit rate. In the systems recommended by the CCITI this is accomplished by having the second order bit rates somewhat higher than four times the ideal pri• mary bit rates, thereby ensuring thatt, even "fast" primary multiplexes can be treated in a proper ways. A second order PCM multiplex is already in use in the United States. This system combines 96 analogue channels to a 6312 kbit/s dig• ital signal, t.e. it uses the same transmission line as the dlgtttal multiplex according to figure 2-8. In Europe a second order PCM multiplex based on the parameters of the 30-channel system and the 8448 kbit/ s trans• mission line has been discussed. This system is proposed to be time slot interleaved with 132 time slots. Of these, 128 can be used for speech channels, 2 for synchronization and 2 for signalling.

2.4.1. TDM Hierarchy Based on 30 Channel System

The digital transmission systems can be tied together in a hier• archy in the same way as the FDM systems. Two hierarchies are dis• cussed, one based on the 24-channel system and the other based on the 30-channel system. One possibility for the latter is shown in figure 2-10. From the figure, it can be seen that the transmission facilities are planned to be used not only by pulse code modulated speech but possible digit rates and ttransmission media are shown in figure 2.11.

31 1st order 2nd order 3rd order 4th order 5th order 2.048 Mbit/s 8.448 Mbit/s 34.368 Mbit/s 139.264 M bit/s -565 Mbit/s VF 1 Primary . I PCM t-- 30 1 Multiplexor Second Order 1 Data.:.-· ----, Digital Data Mulliplexcir -- Multiplexor.,__ ffi --- ...L Third VF Order I-- u, Second Order I'.:\ I Digital ~Multiplexor •. ;] PCM Multiplexor 2 1-=-l Fourth Visual Order 1 Digital ---1 Telephone 1------• ~ Coder ~Multiplexor Fifth FDM 4 Order i-• -----4 Master t-- Digltal . Group ~- Multiplexor

TV 4 --- Coder

Figure 2-10. A possible digital transmission hierarchy based on the 30-channel PCM system.

N N :i: Q) :i: (!) Q) Q) (!) (I) :0 0 Q) :0 ::0 "O 0"' ~ ·5 "'0 "'0 oi "' "'. 0 O> oi oi ·x M Q) ·x ·x Q) > Order Digit rate No. of Q) Q) "'0 Ji "O -a, 0 0 0 C "'3: Mbit/s ::0 .Q a. "' "' ·5 s~eech 0 0 oi - O> oi c annels "'0 o-"'>, o' = E .2 Q) .~ for digital ·l;j .!: 3: •.. "' "O > "'Q) .2 E 0 a. multiplexes e, Q.c ::i: (J) z a:"' ~"' 0

2.048 30 X X

2 8.448 120 X X X X X

3 34.368 480 X

4 139.264 1920 X X X X X

5 -565 7680 X X X

Figure 2-11. Digit rates and transmission media for a hierarchy based on the 30-channel PCM system

32 •

2.4.2. TDM Hierarchy Based on 24-Channel System

To take further advantage of the merits of TDM and digital trans• mission, the common carriers employ a hierarchy of further multiplexing as shown in Figure 2-12. Four'I'I lines are multiplexed in an Ml2 multi• plexer to generate a T2 transmission system, seven T2 lines convert to a

I T3 line in an M23 multiplexer and six T3 lines convert to a T4 line in an M34 multiplexer. At each stage additional frame synchronizing bits must be added as with the first order multiplexing so that at each multiplexer output it would be possible to distinguish which bits belong to which in• put.

1.544 Mb/s s,_..s,--+ .: 6.312 Mb/s • • PCM s,.__.• I 2---.Tl I M12 3~

4~ M23 T2 2--+ 274.176 Mb/s • • ~ • M34 7--+ T2 T3 2--+ • •

Figure 2-12. TDM hiararchy.

33 There is a problem that arises in connection with the higher orders of multiplexing that does not occur at the first order. In the first order multi• plexing there is Just a single clock to contend with, that is the clock that drives the commutator. However, the four input lines in the Ml2 multiplexer come from physically widely separated location and employ four separate un• synchronized clocks.' These clocks are very stable crystal controlled os• cillators which are, of course, set to operate at the same nominal frequency as nearly as possible. Since, however, chronized and will hence experience a relative frequency drift. Design specifications allow a drift from the nominal set frequency of± 130 parts/million. As may be verified, if then two clocks have frequencies which differ by 2 x 130 = 260 parts/million,-the faster clock will have generated one more time slot than the slower clock in the course of Just 20 frames. For proper interleaving of bits at the Ml2 level it is necessary that all input bit streams have or be made to appear to have the same rate. To put the matter most simplistically, the process of adjusting bit rates to make them equal involvesadding bits to the slower bit stream in an operation referred to as "pulse stuffing." Further bits must then be added to all bit streams to allow the receiver of the composite signal to distinguish time slots which carry information from slots which carry the "stuffed" bits. The Ml2 multiplexer adds nominally 17 bits for frame syn• chronization and pulse stuffing. Hence the number of bits per frame is

193 x 4 + 17 = 789 bits/frame

The T2 line bit rate is therefore

fb(TT2)= 789 bits/frame x 8000 frames/s = 6.312 Mb/s

34 •

The M23 multiplexer adds nominally 69 bits for synchronization and pulse stuffing; hence the number of bits per frame for a T3 line is

789 x 7 + 69 = 5592 bits/frame

and fb(T3)= 5592 X 8000 = 44. 736 Mb/s

The M34 multiplexer adds nominally 720 bits for synchronization and pulse stuffing and therefore tthe T4 system has a bit rate

fb(T4)= 274.176 Mb/s

A detailed analysis of the architecture and operation of these digital systems it to be found in Ref.3.

35 ..

CHAPTER III THE TELEPHONE NETWORK

The invention of the telephone in 1876 led to an explosive outgrowth of engineering developments. These developments continue to thrive to this date. This proliferation, through what has become known today as "the larg• est industry in the world." has led to the existing informational era in which we live. Despite the Bell System divestiture (breakup) on January 1, 1983. The telecommunicattions industry sttill exists, employing millions of people. Divestiture has clearly established a distinction between the principal en• tities or segments of the telecommunications industry. - Over 1400 in• dependent telephone companies exist today, making up what has become known as the public switched telephone network (PSTN). The PSTNhas been dubbed the "world's most complex machine." It is truly one of the modern wonders of the world. In the United States alone, the PSTN interconnects hundreds of millions of telephones through the largest network of computers in the world. Despite its electrical and mechanical sophistication, even the most unskilled person can utilize its services. At the touch of a dial, any two people, virttually anywhere in the world, can communicate with each other in a matter of seconds. In this chapter we examine the basic structure of the PSTNand how it has evolved over the years. The most recent technological advances are introduced. 3.1. Th•e Public Switched Telephone Network (PSTN) Since the invention of the telephone, the PSTN has grown pro• portionately with the increased demands to communicate. Switching ser• vices beyond metropolitian areas were soon developed, increasing the size and complexity of the central office. New methods of switching were re• quired to interconnect central offices through the use of interoffice trunks

36 and tandem trunks. Figure 3-1 depicts how today's central offices are con• nected through the use of trunks. In largely populated areas, a tandem of• fice is used to minimize the number of trunks that a call must be routed through to reach its destination. Outside the local area, toll trunks are used to connect the central office to toll centers. Toll centers may be locat• ed in adjacent cttttes outside the local area. To achieve even longer dis• tances, toll centers are interconnected by inttertoll trunks as shown.

/ / Toll frunk / I I / / /Toll center

Toll trunk

Figure 3-1 Interconnection of switching exchanges in North America.

37 3.1.1. Switching Hierarchy of North America

A hierarchy of switching exchanges evolved in North America to ac• commodate the demand for longer-distance connections. The PSTN has classified these exchanges into five levels of switching, as depicted in Table 3-1. At the lowest level, class 5, is the central office or end office. A large metropolis may require several end offices for service, whereas in a rural area a single end office is usually sufficient. When calls are made outside the local area, they are routed through class 4 centers, toll centers, and possibly higher levels of switching, depending on the destination of the call and the current traffic volume within the PSTN. To aid in the volume of traffic between toll centers, primary centers, sectional centers, and re• gional centers, classes 3,2, and 1, respectively, are used. Figure 3-2 il• lustrates the possible routes that a toll can take throughout the switching hierarchy. The best route is the shortest route or the route utilizing the smallest number of switching centers. A call may not always take this route, however. It depends on the availability of trunk circuits when the call is placed. Several alternative routes can be taken up and down the switching hierarchy. The selection of a route is under program control at the switching center.

TABLE3-1. Hierarchy of Switching Exchanges in North America.

Class 1 Regional center D . Class 2 Sectional center 6

Class 3 I Primary center I Q Class 4 I Toll center ,e

Class 5 I End office 1G)!,·

38 .•

Regional I Class Sooa11d choice center 1 Class Regfonal ------·------1 center

Sectional center Sectional center

Primary center

Toll center

Long-dlstance calf from party A to party B (no dire.ct lntemlilcc trunk) A B

Figure 3-2. Long-distance switching hierarchy within the PSTN de• picting plossible routes to complete a call between parties A and B.

Typically, tru~s used to interconnect higher levels of switching are designed for high-speed transmission using the multiplexing tech• niques that will be described. These higher levels of switching make up the long-haul network. Calls placed onto the long-haul network are sub• ject to toll fees. Figure 3-3 illustrates the long-haul network. To sustain higher transmission rates within the long-haul network, wideband trunk media are used. This includes coax, microwave ground and satellite links,

and fiber-optic cables.

-- Microwave radio - - - Cable ,__ Satellite • No through traffic o Through traffic only • Through and terminal traffic Junction office-through and terminal traffic

Figure 3-3. Long-haul network depicting various trunk media. (Cour•

tesy of Bell Laboratories.)

3.2. Technical Consideratons in the Planning of Digital Net• works for Multiexchange Areas

The use of remote concentrators to connect subscribers will make it economical to increase the exchange areas. Further the digital group se• lectors will have a capacity to handle the traffic from a large number of

40 .• subcribers. The combination of digital group selectors and common chan• nel signalling will make the use of both way circuits between the ex• changes more attractive. Consequently the digital network will contain larger exchanges with more direct routes and there will be fewer tandems in the local network. But the same routing alternatives remain; direct routes or tandem connections. Furthermore, as the local network only contains a two-level exchange hierarchy and in the digital case there are fewer exchanges (tandems) on the highest level, fairly simple introduction strategies can be outlined. It also appears that there is little difference be• tween overlaying and replacement when located exchanges are concerned. A convenient introduction strategy could be to install a dtgtttal tan• dem at an early stage and let that tandem take care of connuctions be• tween the analogue and the digital network. Such a solution would be in favour of the final network structure and it would support the new tech• nique and help avoid further development and extension of existing an• alogue exchanges. When the old parts are replaced the "transfer" capacity of the tandem exchange will be used for the growing "normal" tandem traffic. If this increases slowly, the "overcapacity" in the tandem exchange may be used for concentrators connected directly to the tandem exchange. To what extent this method is applicable in practice is of course dependent on transmission and signalling in the network, and above all it depends on cost factors. Therefor, firm conclusions could only be drawn after a thor• ough study of the actual case.

3.2.1. Numbering

Together with the planning of the network structure a long-term numbering plan should be made up and number series be allotted to new

41 exchanges in accordance with that plan. In order to avoid unnecessary confusion one should adopt the principle that concentrators be included in the number series of their parent digital local exchange. AB a matter of fact it is most likely that all the subscriber subsystems, remote con• centrators included, that are connected to the same digital group selector will be contained in one common free number group. Sometimes an area will have enough subscribers only for some concentrators but no for an own group selector. The concentrators will be connected to a group selector in another area until it is economical to in• troduce the group selector. In such a case the best solution is to include the concentrators in a number series planned for the future parent digital group selector.

3.2.2. Transmission Systems

AB mentioned in chapter 2 the European PCM hierarchy will prob• ably be 2, 8, 34, 140 and 565 Mbit/s (30, 120, 480, 1920 and 7680 chan• nels}. The 2 Mbit/s transmission is conveyed on the existing type of pair• cables, while special cables (shielded pair or coaxial cable) are required for the higher systems. Radio link transmission is not of interest in urban ar• eas. Of special interest is the 140 Mbit/s transmission on small diameter coaxial cable. Between analogue exchanges it is normally profitable to use PCM on routes which are longer than 6-12 km, while between analogue and digital or between digital exchanges it is profitable to use digital trans• mission regardless of the distance. The digital transmission network is built up of digital sections which are common to many routes of various size and types. The capacity

42 .. of the section is consequently much larger than that of the routes. In the beginning the digital sections will be combined to a star-type network, but later on a change towards a mes-htype network will occur. The larger metropolitan areas may in the future consist of a mesh• type network of 140 Mbit/ s systems between the central large exchanges and a star-type network of 2 Mbit/s systems to the pertfertcal exchanges.

3.2.3. Transmission Planning

The reference equivalent is the most important parameter for transmission planning. CCITI recommends not only upper limits but also mean value for the national reference equivalents. The mean value of the sending reference equivalent (SRE) should be between 10 and 13 dB and the receiving reference equivalent (R-RE)should be between 2.5 and 4.5 dB. The mean overall reference equivalent fo international connections will then be 13-18 dB. For the transmission planning in a network developing according to the overlay principle the followingapplies. With the introduction of PCM on the longest routes between the local exchanges, the reference equiv• alents for the longest connections will be reduced. The further develop• ment with digital switching and transmission will give more subscribers a lower value on SRE and RRE. When digital telephone sets are introduced, the reference equivalents for these can be chosen to a suitable value giving a more optimal value for the distribution of reference equivalents. The "ful• ly digital goal" when all switching and transmission is dgttal will result in all subscribers having identical values of SRE and RRE. The transmission planning will thus consist of among other things, planning the distribution towards an ideal value which also includes the avoidance of too low val-

43 ues for the reference equivalents.

The replacement of an analogue exchange with a digital one, while keeping the analogue transmission links, is equivalent to an introduction of an extra transmission link. If the analogue links are two-wire, the di• gtttal exchange will from a small four-wire loop, where the stability has to be mentioned. One way to do this is to insert build-out pads, which re• sults in the distribution of the reference equivalents being changed to higher values. On the other hand the transmission links will be extended by or replaced with digital links. Then the development will be similar to that sketched for the overlay principle.

3.2.4. Signalling

A digital network, overlaid or replacing the existing analogue net• work, should use common channel signalling (CCS). The amount of signalling equipment for interworking with the ex• isting analogue network is minimized if the 'interface between the analogue and the digital networks is not crossed more than once for a particular call. The amount of signalling equipment is also reduced if the inter• working can be carried out at higher levels in the network. This means a better use of the equipment and also makes it possible to completely avoid MFC-signallingin digital exchanges at lower levels. On a PCM-route between a digital exchange and an analogue SPC• exchange, the same CCS system as in the digital network should be used. In case of analogue transmission links, an analogue version of the CCS system can be used, or a quasi-associated mode of operation can be em• ployed. On a route between a digital exchange and non-SPC analogue ex-

44 change, PCM is prefered together with a proper type of channel associated signalling.

The signalling system between the digital local exchange and a con• centrator is largely dependent on the switching system. It may be based on thesame principles as used in the CCS-system for autonomous exchanges.

Subscriber Network

r--- 1 I

, "' I Digital I 2.X£h,!n_g~

Junction Network r------~--, I I

I I I I I Old exchange : 1 Digital group selector 1 building · Q L ------.J . 0 Concentrator -D Cabinets DiQital transmi!;!;inn

Figure 3-4 Different locations of concentrators

45 3.3. Network Structure in the Rural Area

The information presented so far has treated solutions for urban areas. There we will have large switching points and, due to the hier• archical structure and alternative routing, a high utilisation of the trans• mission media. For the long distance network the sittuation is similar, with a slight difference in the introduction of digital technology depending on the different transmission medias used existing SPC-exchanges etc. An area asking for quite different solutions is the rural area. Due to large distances it has been difficult to find economically feasible solu• tions with existing centralized switching. This, together with other factors, has in many countries resulted in a lower telephone penetration in these areas. The need for decentralized switching may be met by distributed concentrator systems controlled from a group centre and, if economical, with internal switching facilities. The low utilization of transmission media due to small groups can be overcome by a common use of the PCM-system by different con• centrators. All this will give us a new structure of the rural network with remote controlled concentrators linked together by a common PCM sys• tem. The network will have a good utilization of transmission media and group selector inlets as well as plossibilities of centtralization of common control functions. Different examples of structures are given in Figure 3.5.

46 To/from the LD Network

• Group selector

'CJ Concentrator

PCM-transmission

Populated areas

Figure 3.5. Different structures for a rural network. witth con• centrators.

3.4. A Completely Digital Network

The trends described above should ultimately lead to a· completely digital network. This implies the penetration of digital transmission down

47 to the individual subcriber set, where analogue/ digital conversion is per• formed. As mention abov ethis penetration is a long-term one. Apart from the technical and economical problems, an international standardization is desirable before any large-scale introduction of digital telepbhone can take place. A model of a completely digital network, applied to a multi• exchange area, is shown in Figure 3-6.

To/from the trunk network

Trunk/tandem exchange

Local exchange group selector

Concentrator

Digital transmission line

Figure 3.6. A completely digital multiexchange network. Analogue/ digital conversion is performed in the subscriber set.

TThe network consists of large local exchange group selectors con• nected in a mesh structure, controlling a number of digital concetrators, which may be remotely located or colocated with the group selectors. The concentrators together with their group selector from a star structure. Ac• cess to the trunk network is arranged by a digital trunk exchange. At this point a tandem exchange may be situated, switching overflow traffic be• tween the local group selectors. All transmission is digital. Analogue/

48 • digital conversion, using for example PCM or delta modulation, is per• formed on a per channel basis in the subscriber set. Very much has been left out of the picture. The service integration of telephony, telex and data has not been considered, neitther has the in• fluence caused by visual telephone, cable 1V etc. Tihis total picture can• not be foreseen for the present.

49 CONCLUSION

The project provides a strong comprehensive and illuminating a knowledgement about PCM technique and itts application in digital teleph• ony. The outors have tried to present in detail basic operations of PCM technique: saampling, quantizing, compaunding and coding. The project can be used by students in the laboratory of "Tele• communication" and by engineers in design PCM system and telephone network.

50 REFERENCES

1. John G. Proakis. Masoud Salehi. Communication System En• gineering, Prentice-Hall, 1993. 2. Bruce Carlson. Communication System. McGraw-HillCo., 1986. 3. Kennedy Davis. Electronic Communication System. McGraw-HillCo. 1992,

4. Fred Taylor. Principles of Signal & Systems. Mc Graw-HillCo. 1994. 5. Martin C. Roden. Analogue &Digital Communication-Systems. Pren- tice-Hall, 1992. 6. Warren Hioki. Telecommunication. Prentice-Hall, 1995. 7. . Jack Quinn. Digital Data Communications. Prentice-Hall, 1992. 8. Taub Shilling. Principle of Communication Systems. McGraw-Hill Co., 1992. 9. Santa Clara, Digital, Introduction to LocalArea Networks, Digital Equipment Corp, CA, 1982 10. Fike, J.L. and Friend, G.E., Understanding Telephone Electronics, Texas Instruments, Dallas, Texas, 1990.